Nonlinear-elastic fracture mechanics methods are used to assess the fracture toughness of bulk metallic glass ͑BMG͒ composites; results are compared with similar measurements for other monolithic and composite BMG alloys. Mechanistically, plastic shielding gives rise to characteristic resistance-curve behavior where the fracture resistance increases with crack extension. Specifically, confinement of damage by second-phase dendrites is shown to result in enhancement of the toughness by nearly an order of magnitude relative to unreinforced glass. © 2009 American Institute of Physics. ͓DOI: 10.1063/1.3156026͔The absence of microstructure in monolithic bulk metallic glasses ͑BMGs͒ can lead to marked strain localization and rapid shear-band propagation. 1,2 This effect, which causes extremely low macroscopic plastic deformability, can be devastating to mechanical performance in that properties that are limited by the extension of cracks, such as tensile ductility, toughness, and fatigue resistance, can become severely compromised. Specifically, unstable fracture can ensue in monolithic BMGs along a single shear band with essentially zero macroscopic ductility, 3,4 such that the toughness is far lower than in comparable crystalline alloys. The essential element to developing high toughness in BMGs is to prevent single shear-band failures. By introducing a second phase in form of crystalline dendrites and by matching microstructural length scales ͑i.e., interdendritic spacing͒ to the mechanical length scales ͑i.e., critical crack size for failure͒, nonlocalized plasticity in metallic glasses can be enhanced significantly. [5][6][7] Indeed, the recent development of in situ bulk metallic glass-matrix composites has shown that provided this phase acts to arrest shear-band propagation over appropriate size scales, 8-10 the problems of poor ductility, toughness, and fatigue resistance can be mitigated.One problem here is that the new composite BMGs are undoubtedly far tougher than monolithic BMGs or some earlier BMG composites and current processing methods often cannot make section sizes large enough to meet the critical validity requirements for accurate fracture mechanics measurements, i.e., not meeting fracture mechanics requirements for valid stress intensity K-or J-dominated crack-tip fields and/or for plane-strain constraint. Indeed, many toughness measurements on BMG materials reported in the literature [11][12][13][14][15] are inaccurate due to problems of inappropriate measurement techniques ͑e.g., the area under a compression stress/strain curve͒, absence of sharp stress concentrators ͑e.g., using a relatively blunt notch rather than a fatigue precrack͒ and insufficient test-sample size. Moreover, while single-value measurements, such as K Ic , properly define the toughness of nominally brittle materials, they can be insufficient for alloys displaying extensive plastic deformation and subcritical cracking, as can occur in many BMG composites. Stable crack growth in metallic glasses is not generally observed and ...